25 Robotics Startups to Watch in 2026, According to Investors

Nominated by: Amy Yin, defy.vc (an investor)

Total funding: $6.5 million, according to the company

What it does: San Francisco-based Avatar Robotics uses remotely operated robots to sort and pack products in warehouses. The company uses data from those shifts to train its robots to become more autonomous.

Why it’s promising: Avatar’s robots generate training data while doing paid work for customers. Human operators in Mexico and the Philippines control them with virtual reality headsets and handheld controllers, and Avatar uses those sessions to improve its robotics models.

CEO Colin Webb previously told Business Insider that this approach produces more useful data than paying people to collect it separately. His long-term goal is to build an “unlimited workforce for industrial labor” and have millions of robots deployed across factories, farms, and mines.

Nominated by: Ben Hemani, Bison Ventures (an investor)

Total funding: $160 million, according to the company

What it does: Santa Clara, California-based Cobot builds Proxie, a mobile robot that works alongside people in hospitals, factories, and logistics centers to move carts and materials. Its latest version has two arms for carrying and restocking items.

Why it’s promising: Cobot is led by Brad Porter, who previously ran Amazon Robotics and served as CTO of Scale AI. Its first-generation robots logged nearly 13,000 hours and moved more than 154,000 carts for customers, including Mayo Clinic and Maersk, Porter said.

Those deployments shaped a second-generation Proxie that is more compact, can lift more than 200 pounds, and is cheaper to build because it has fewer parts.

Hemani of Bison Ventures told Business Insider that Proxie’s autonomy stands out. A worker can, for example, write a cart’s destination on an attached whiteboard, and Proxie can read it and deliver the cart there autonomously.

Nominated by: Mark Martin, Cybernetix Ventures (an investor)

Total funding: $47.5 million, according to the company

What it does: Boston-based Dash Bio automates lab testing for drug development. Its robots and software handle work that scientists do manually, such as preparing samples and running tests, helping pharmaceutical companies deliver results faster.

Why it’s promising: Dash Bio was cofounded by Dave Johnson, Moderna’s former chief data and AI officer. Dash can complete testing in days, around 10 times as fast as industry norms, Johnson told Business Insider. The company has a few dozen customers and publishes fixed pricing online, avoiding a lengthy sales process.

Nominated by: Kahini Shah, Obvious Ventures (an investor)

Total funding: $295 million, according to the company

What it does: Bay Area-based Dexterity is building a general-purpose AI “brain” designed to control many kinds of robots.

Why it’s promising: Dexterity’s flagship product, Mech, is an “industrial superhumanoid,” CEO Samir Menon told Business Insider. It has large arms and an extended reach for tasks such as loading trucks and stacking boxes onto pallets. Mech can load roughly 400 boxes into a trailer, compared with an industry average of about 300. The robots are powered by Foresight, Dexterity’s AI model, which decides in real time how best to arrange packages inside a truck.

FedEx recently expanded its deployment of Dexterity’s robots at its Maryland hub after several years of testing. Menon said Dexterity’s systems are deployed at around 10 sites, with a similar number of pilots underway.

Nominated by: Alexandra Sukin, Bessemer Venture Partners (not an investor)

Total funding: $405 million, according to the company

What it does: Founded in 2023, FieldAI is building a single AI system that can control different types of robots and help them operate safely and autonomously in unpredictable environments.

Why it’s promising: FieldAI has more than 30 customers, including construction firms, data center operators, and defense companies. CEO Ali Agha told Business Insider in June that the company has secured more than $100 million in revenue and customer contracts.

The company’s software evaluates risk and makes decisions in unfamiliar situations without extensive task-specific training. The company has recruited from Google DeepMind, Waymo, Tesla, Nvidia, Boston Dynamics, and Amazon.

Nominated by: Alexandra Sukin, Bessemer Venture Partners (an investor)

Total funding: $60 million, according to the company

What it does: San Francisco-based Foxglove builds software for robotics companies, helping teams analyze data from its robots, identify failures, and improve performance.

Why it’s promising: CEO Adrian Macneil got the idea for Foxglove while leading infrastructure engineering at Cruise, where he said less than 1% of the data collected across millions of miles was useful for improving its self-driving system.

“I realized that every robotics company would eventually need the same foundational data infrastructure, and they shouldn’t have to spend years building it themselves,” Macneil told Business Insider.

Foxglove now has more than 250 customers, including autonomous driving startup Wayve and defense tech companies Anduril, Shield AI, and Saronic.

Nominated by: Siddharth Ramakrishnan, Scale Venture Partners (an investor)

Total funding: $540 million, according to the company

What it does: Bay Area-based Generalist builds AI software that can control different robots across factories, laboratories, and homes.

Why it’s promising: Former Google DeepMind and Boston Dynamics roboticists founded Generalist. To teach its AI about the physical world, the startup built low-cost “grippers” that mimic robot hands and distributed thousands of them around the world, collecting more than 500,000 hours of training data.

CEO Pete Florence told Business Insider that the approach is working. He said Generalist’s first model, GEN-0, improved as the company added more data and computing power. Florence compared its latest model, GEN-1, to GPT-3 because some uses, including folding laundry and packing iPhones into boxes, are now becoming commercially viable.

Nominated by: Mark Martin, Cybernetix Ventures (not an investor)

Total funding: $347 million, according to the company

What it does: Pittsburgh-based Gecko Robotics builds wall-climbing robots and AI software to inspect power plants, pipelines, and naval ships.

Why it’s promising: CEO Jake Loosararian started Gecko after visiting a power plant in college and learning how dangerous manual inspections could be for humans.

“If you want to digitize the physical world, you have to go about the hard effort of collecting data that’s never been digitized before, and at a cost that isn’t prohibitive,” Loosararian told Business Insider. “That’s where robots come in.”

The company has since expanded to software that helps customers plan repairs and maintenance. In March, Gecko won a five-year contract with the US Navy and General Services Administration worth up to $71 million, starting with inspections of 18 Pacific Fleet ships.

Nominated by: Ryan Gembala, Pathbreaker Ventures (an investor)

Total funding: $86 million, according to the company

What it does: Carson, California-based GrayMatter Robotics builds AI-powered robots for surface finishing, a process that includes sanding, grinding, blasting, and coating manufactured parts. Its robots work on products ranging from guitars to fighter jets.

Why it’s promising: Traditional industrial robots are typically programmed for a specific part and struggle when its shape, material, or condition changes. GrayMatter’s robots scan each part, adapt on their own, and learn from every deployment. The company says its systems have generated more than 7 petabytes of manufacturing data, which it uses to improve its AI.

GrayMatter’s customers include the US Air Force, Boeing, and Lockheed Martin. This month, the company announced a seven-year, $900 million agreement with two other robotics companies to automate shipbuilding work for the US Navy.

Nominated by: Ryan Gembala, Pathbreaker Ventures (not an investor)

Total funding: $169 million, according to the company

What it does: Los Angeles-based Machina Labs builds robotic factories to make complex metal parts. Its RoboCraftsman system uses two robotic arms, sensors, and AI to shape, inspect, trim, and drill parts directly from digital designs.

Why it’s promising: Ed Mehr, who previously worked at SpaceX, leads Machina. The company builds what Mehr calls a “software-defined factory” that can switch between making different products without costly retooling.

“You can program that facility to make missile components in the morning, and then in the afternoon you make aircraft components, and then at night when nobody’s there you make automotive components,” Mehr told Business Insider last month.

Machina is building a 200,000-square-foot factory with room for 50 RoboCraftsman cells. Last month, the company won a contract with Lockheed Martin to qualify its manufacturing process for parts used in JASSM, a long-range cruise missile used by the US military.

Nominated by: Charly Mwangi, Eclipse (an investor and cofounder)

Total funding: Undisclosed

What it does: Bay Area-based Matter is building autonomous factories that assemble drones and robots.

Why it’s promising: Matter was cofounded by a team with experience at Tesla, Rivian, and Travis Kalanick’s Atoms, and is being incubated by Eclipse partner Charly Mwangi, who previously held executive roles at all three companies. The startup has opened a 54,000-square-foot factory in Sunnyvale, California, and begun producing robotic components for customers.

“The goal is to rebuild American manufacturing as a software problem, not a labor problem,” CEO Adi Prasad told Business Insider last month.

Matter says each production run generates data that improves MatterOS, its platform for automating and coordinating factory operations.

Nominated by: Charly Mwangi, Eclipse (not an investor)

Total funding: $68 million, according to the company

What it does: New York City-based Mecka AI collects training data for robotics companies by capturing how humans interact with the physical world. People record everyday tasks using phones and wearable devices, and Mecka turns the recordings into datasets for training robots.

Why it’s promising: Mecka aims to make physical-world data as inexpensive and easy to collect as the text and images used to train traditional AI models.

Instead of relying on people to control robots, Mecka puts sensors on humans to capture the visual, movement, and force data that cameras miss. It works with leading AI labs and is building custom hardware, including gloves with built-in sensors, for data collection. Mecka recently crossed $100 million in annualized revenue, CEO Josh Gao told Business Insider.

Nominated by: Ben Hemani, Bison Ventures (not an investor)

Total funding: $20 million, according to the company

What it does: Zurich-based Mimic Robotics builds dexterous robotic hands and AI for complex industrial tasks. Workers use wearable gloves to demonstrate tasks, and Mimic’s software trains robots to replicate those movements.

Why it’s promising: Mimic spun out of an ETH Zurich research lab two years ago. The startup is working with manufacturers, including Audi, which is testing its technology in factories. The company says that its five-fingered robotic hand, which can handle more than 55 pounds, is not far off human capabilities in speed, dexterity, and strength. The startup is one of several trying to solve what Elon Musk has called the “hands problem” — recreating the dexterity of the human hand with sensors, motors, and software.

Nominated by: Ally Warson, UP.Partners (an investor)

Total funding: Undisclosed

What it does: Bay Area-based Noble Machines builds general-purpose humanoid robots for hazardous and physically demanding industrial work. Its Moby robot is designed for jobs in manufacturing, construction, and energy plants.

Why it’s promising: Engineers from Apple, SpaceX, and NASA founded Noble Machines in 2024. In less than two years, the company has built three generations of its humanoid robot and is finishing a fourth.

CEO Wei Ding told Business Insider that Noble has a deal to deploy its humanoids with a major global construction company. The startup is also developing AI that could eventually let customers teach a robot a new skill as quickly as it takes to train a new employee.

To reduce geopolitical risk and compete globally, Noble maintains two supply chains for its humanoids: one in Asia and another across the US and allied countries.

Nominated by: Lyna Kim, Lowercarbon Capital (invested personally)

Total funding: Undisclosed

What it does: San Francisco-based Palatial automates the creation of virtual environments where robots can practice tasks before attempting them in the real world.

Why it’s promising: Steven Ren got the idea while working on a digital twin of Tesla’s Nevada Gigafactory, a virtual model used to plan and coordinate construction. Building it was slow and manual, and Ren saw an opportunity to automate the process. He soon concluded that robotics was an even bigger market than construction — a building needs one digital twin, while training a robot requires millions of simulated objects and environments.

Since launching its automated simulation generator earlier this year, Ren told Business Insider that more than 200 AI companies have expressed interest, with many — including Nvidia — becoming customers.

Nominated by: Stephanie Zhan, Sequoia (an investor); James Detweiler, Felicis (an investor)

Total funding: $2.2 billion, according to PitchBook

What it does: Skild AI builds an AI model designed to control robotic arms, humanoids, quadrupeds, and mobile robots. The model is designed to help different machines learn new jobs without being programmed for each task.

Why it’s promising: Former Carnegie Mellon professors Deepak Pathak and Abhinav Gupta founded Skild. The company generated $30 million in revenue in 2025. Its technology is used in security, construction, warehouses, and manufacturing. Skild is working with Nvidia and Foxconn to use its AI to control robotic arms on production lines for Nvidia’s Blackwell chips, automating an assembly task that is currently performed by people.

Nominated by: Amanda Huang, Bain Capital Ventures (an investor)

Total funding: $200 million, according to the company

What it does: Bay Area-based Sunday Robotics builds Memo, a wheeled robot that performs household chores.

Why it’s promising: Last month, Sunday said Memo, powered by its new AI model, folded laundry in unfamiliar homes with a success rate above 99%.

“I think 2026 will be when the timeline for autonomous home robots shrinks in a very drastic way because of this breakthrough,” CEO and cofounder Tony Zhao, a former Stanford robotics researcher, told Business Insider last month.

Sunday plans to place Memo in homes through a beta program this fall. It is also training Memo to vacuum, organize toys, and make coffee.

Nominated by: Lyna Kim, Lowercarbon Capital (an investor)

Total funding: $8 million, according to the company

What it does: Bay Area-based Synphony collects real-world data and builds software that adapts robotics foundation models to specific industrial jobs. Its first challenge: teaching robots to pick strawberries.

Why it’s promising: Synphony participated in Y Combinator earlier this year and finished the program with more than $1 million in revenue, CEO and cofounder Sean Wu told Business Insider. For its first pilot, the startup put data-collection gloves on strawberry pickers at a farm run by one of the largest US producers, then used the data to train a robot to pick the fruit.

Synphony plans to expand into other industrial tasks, including meat processing and warehouse packing.

Nominated by: Stephanie Zhan, Sequoia (not an investor)

Total funding: $265 million, according to PitchBook

What it does: Hangzhou, China-based Unitree builds humanoid and four-legged robots for research, consumer, and industrial uses.

Why it’s promising: Founded in 2016 by CEO Wang Xingxing, Unitree has shown that humanoid robots can be built relatively cheaply and at scale. It shipped more than 5,500 humanoids in 2025, more than a quarter of the global market, according to the company. Unitree is preparing to go public in Shanghai at a valuation of about $9 billion, potentially making it China’s first publicly listed humanoid robotics company. It plans to invest in developing new hardware and software and expanding manufacturing.

Nominated by: Siddharth Ramakrishnan, Scale Venture Partners (not an investor)

Total funding: $7.5 million, according to PitchBook

What it does: Waterloo, Canada-based Upside Robotics builds small, solar-powered robots that apply fertilizer directly to plants. They use sensors and software to determine how much fertilizer each plant needs rather than applying the same amount everywhere.

Why it’s promising: The company says its robots can cut fertilizer use by up to 70% by delivering nutrients directly to plant roots. Built-in cameras and sensors also monitor crop health.

Nominated by: Jim Adler, Toyota Ventures (an investor)

Total funding: $300 million, according to the company

What it does: Cambridge, Massachusetts-based Walden Robotics, which spun out of the Toyota Research Institute, builds general-purpose robots and the AI models that control them.

Why it’s promising: Walden deploys robots in factories and warehouses to perform repetitive, physically demanding tasks such as loading car parts, cleaning machinery, and preparing components for assembly. CEO Russ Tedrake told Business Insider that the robots move on wheels rather than legs for greater safety and stability, and to leave room for a larger battery and more computing power.

“It’s a pragmatic version of a humanoid,” said Tedrake, who is a longtime MIT roboticist whose former students hold leadership roles at Boston Dynamics, Google DeepMind, Generalist, and Sunday Robotics.

Adler said he’s bullish on Walden because of Tedrake’s experience and the company’s early traction. Within months of launching in January, its robots were handling eight-hour shifts at customer sites, he said

Nominated by: Jim Adler, Toyota Ventures (not an investor)

Total funding: $11 million, according to the company

What it does: South San Francisco-based Westmag, short for Western Magnetics Company, makes motors and actuators for drones and robots.

Why it’s promising: China dominates production of many of these components, but Westmag, with a seed round led by Andreessen Horowitz, is developing a US-made alternative to reduce manufacturers’ reliance on overseas suppliers. The company is now ramping up production at its first factory south of San Francisco.

“The Western drone and robot companies we serve need reliable, cost-effective, and resilient component supply in order to scale and win globally,” Jordan Sanders, cofounder and COO of Westmag, told Business Insider.

Nominated by: Amanda Huang, Bain Capital Ventures (not an investor); James Detweiler, Felicis (not an investor); Kahini Shah, Obvious Ventures (invested personally)

Total funding: $78.33 million, according to PitchBook

What it does: XDOF collects and organizes real-world data used for training robots. It gathers data from people who control robots and perform everyday tasks.

Why it’s promising: XDOF is tackling one of robotics’ biggest bottlenecks: getting enough high-quality, real-world training data. Cofounder and CEO Philipp Wu founded the company after encountering a shortage of training data while researching robot learning at the University of California, Berkeley. XDOF now works with about 20 customers, including leading AI labs. Its investors include Thrive Capital, Spark Capital, and Andreessen Horowitz.

Huang from Bain Capital Ventures said that XDOF stands out because it understands that the hardware used to collect training data — from cameras to rig design to tracking algorithms — directly affects data quality. “Garbage in, garbage out,” she said.

Nominated by: Ally Warson, UP.Partners (not an investor)

Total funding: $2 billion, according to PitchBook

What it does: Bay Area-based Zipline builds and operates autonomous delivery drones that carry medical supplies, food, and retail goods.

Why it’s promising: Zipline, which launched in Rwanda in 2016 to deliver blood, is now expanding in the US. The company said that the number of businesses offering delivery through its service grew 13 times in the first half of this year. Now, Zipline has hired executives from Tesla, Waymo, and Uber to scale its business.

Nominated by: Amy Yin, defy.vc (not an investor)

Total funding: $11.05 million, according to the company

What it does: Bay Area-based 1Robot builds software that coordinates people, robots, autonomous vehicles, and AI agents to complete real-world tasks. For example, its software could use camera footage to detect a collision, have an AI agent assess the scene, and dispatch an autonomous vehicle with the right robots and personnel.

Why it’s promising: 1Robot’s customers include Samsung, Nestlé, and government agencies, cofounder and co-CEO Jesse Tran told Business Insider. Revenue reached $5.5 million in the first half of 2026, up from $2.5 million in all of 2025, he said.

Yin from defy.vc said 1Robot also stands out for its Vietnam- and Taiwan-based supply chain, which gives US customers an alternative to Chinese robotics hardware.

“Our future wars are going to be fought with humanoid robotics,” she said.